基于小波变换和尺度共生矩阵的超声图像分割

A Method for Image-Segmentation Based on Wavelet Transform and Scale Co-occurrence Matrix

超声图像所固有的特性使得图像分割比较困难,尤其是应用计算机实现超声图像的自动分割技术远不能达到实际需要,因此,超声图像的分割是亟待解决的一个难题。文中将图像树型框架小波变换、尺度共生矩阵和自组织神经网络聚类相结合应用于超声图像提出一种分割方法。实验表明,应用所提出的方法可得到比较清晰的分割结果,显著提高分割图像的对比度。

Due to the characteristics of ultrasonic image, ultrasonic image segmentation is more difficult, especially to implement ultrasonic image segmentation automatically. It is still an urgent and tough problem to implement ultrasonic image segmentation effectively. This paper proposes a new image segmentation method, which integrates the theory of tree-structured frame-wavelet transform, scale matrix （SCM）, and self-organizing neural network and applies them to the clinical ultrasonic image finish segmentation. Experiment results show a clearer segmented image with a high contrast can be obtained with the proposed method.